> "When I'm starting on a new algorithm, I sometimes find it's both fun and instructional to begin by trying to see just how simple-minded an algorithm I can wr

> "When I'm starting on a new algorithm, I sometimes find it's both fun and instructional to begin by trying to see just how simple-minded an algorithm I can write. I think I do this just to prove to myself that the problem can at least be solved. From there, anything else is mere refinement." This is excellent advice.